df1412/anime-big-lama

AnimeLaMa Inpainting Model

TorchScript model for anime/manga image inpainting - removes text and reconstructs background.

Model Details

• Model: AnimeLaMa (Big LaMa finetuned on anime/manga)
• Format: TorchScript (.pt)
• Size: ~197MB
• Input: RGB image + binary mask (white = inpaint, black = preserve)
• Output: Inpainted RGB image

Usage

import torch
from huggingface_hub import hf_hub_download

# Download model
model_path = hf_hub_download(
    repo_id="deckyfx/anime-big-lama",
    filename="anime-manga-big-lama.pt"
)

# Load model
device = torch.device("cpu")
model = torch.jit.load(model_path, map_location=device)

# Inpaint (example)
# output = model(image_tensor, mask_tensor)

Attribution

This model is reuploaded from PanelPachi (08/02/2026) for easier distribution and integration.

Original model is based on Big LaMa finetuned on anime/manga datasets.

License

MIT License - Same as original PanelPachi project.

Credits

• PanelPachi: Original source of this model file
• Big LaMa: Base inpainting architecture
• Reuploaded by: deckyfx